The incidence of breast carcinoma–related death has been steadily decreasing over the past 8–10 years.1, 2 Early detection, effective systemic therapy, multidisciplinary approaches, and advances in supportive care all have received credit for this reduction in incidence. As a result of these efforts, several findings demonstrating the potential curability of breast carcinoma have been made: 1) adjuvant chemotherapy with or without hormonal therapy has been shown to increase DFS and OS; 2) many chemotherapy regimens for the treatment of metastatic breast carcinoma have been credited with improving survival; 3) long-term and durable complete remission of metastatic breast carcinoma has been found to be achievable; and 4) neoadjuvant chemotherapy for early breast carcinoma has been demonstrated to significantly increase pathologic complete remission rates and, consequently, overall survival.

Despite such potential benefits, adjuvant therapy (chemotherapy and/or radiation therapy) may result in numerous delayed side effects that can be associated with significant morbidity; among these side effects are the induction of irreversible premature menopause (resulting in decreased quality of life) and the emergence of chemotherapy- and/or radiation-induced secondary acute leukemia, which is associated with poor prognosis.3 Such secondary leukemia must be distinguished from the classical types of leukemia (French–American–British classification), which may also affect patients with breast carcinoma, albeit very infrequently.4

Numerous reports have described, in patients with breast carcinoma, the occurrence of secondary leukemias that possess specific cytogenetic abnormalities, do not respond to otherwise established and effective treatments, and are associated with very poor prognosis. Several chemotherapeutic agents that are considered leukemogenic, such as alkylating agents, topoisomerase II agents, and podophyllotoxins,3, 5–8 are used in the treatment of breast carcinoma. Radiation therapy9 may also lead to increases in the incidence of secondary leukemia, particularly in patients who receive alkylating agents (e.g., cyclophosphamide) as part of their chemotherapy regimens.10, 11

In the current edition of Cancer, Kaplan et al.12 address the issue of secondary leukemia in patients with breast carcinoma. These investigators report their experience with 2866 patients who had nonmetastatic breast carcinoma and who were identified via the tumor registry of a large community hospital between 1992 and 1999, with follow-up extending to 2001. Approximately 50% of all patients identified were treated for Stage I disease. Adjuvant treatment consisted of radiation therapy, which was received by 85% of patients (n = 2233), and chemotherapy, which was received by 45% of patients (n = 1289); the majority of chemotherapy recipients were treated with anthracycline-based (n = 866) or cyclophosphamide-based regimens (n = 1285; 419 cases without anthracyclines). Patients with leukemia were identified from the registry database, via mail contacts, and by crosschecking with the database of the National Cancer Institute Surveillance, Epidemiology, and End Results Program. Kaplan et al. compared the incidence of leukemia within the study cohort with national age-specific incidence rates and were unable to confirm an increase in incidence among the study population; thus, they concluded that chemotherapy agents, including doxorubicin, were not associated with increased incidence of posttreatment (secondary) leukemia.

Despite this finding, numerous studies, case study reports, and reviews have suggested that there is a chemotherapy-induced increase in the incidence of leukemia among patients with breast carcinoma. Such chemotherapy-induced leukemias have specific cytogenetic features that are not known to be associated with classical leukemia (French–American–British classification); among these features are loss of part or all of chromosome 5 or 7,13, 14 t(11;11)(p13–p15;q23),15 t(3;11)(q28;q23),16 and translocation of the myeloid lymphoid leukemia (MLL) gene, located at 11q23, to chromosome 9, 19, or 4.14 The cumulative cyclophosphamide doses used (> 6 g/m2 via infusion or > 20 mg/m2 orally), particularly when used in conjunction with radiation, are associated with significant increases in the incidence of secondary leukemia. The median time to development of leukemia secondary to topoisomerase II is 2 years,17 and the median time to development of leukemia secondary to radiation therapy plus cyclophosphamide is 66 months.10

Although the study conducted by Kaplan et al. involved a large patient population, their findings did not agree with previously published data. For instance, they did not observe an increased risk of leukemia in association with doxorubicin and/or cyclophosphamide, even when these agents were used in conjunction with radiation therapy. The study included patients with breast carcinoma who were registered between 1992 and 1999, with follow-up extending to 2001 to ensure a minimum follow-up duration of 24 months for every patient. Only 30% of all patients received anthracycline-based therapy, and 46% of patients received cyclophosphamide-based therapy. Because these two agents may induce secondary leukemias that harbor differing cytogenetic abnormalities, they should be analyzed separately over an adequate follow-up period. In addition, because radiation therapy administered in conjunction with cyclophosphamide has been reported to significantly increase the risk of developing secondary leukemia, all patients who received radiation therapy plus cyclophosphamide should also be analyzed separately. Another point to be considered is that 40% of all patients were registered between 1997 and 1999. Thus, the median follow-up duration for these patients may be shorter than the expected time to development of secondary leukemia. The median time to development of topoisomerase II–induced leukemia is 2 years, but the median time to development of alkylating agent–induced leukemia is 5.7 years; furthermore, the possibility of developing secondary leukemia may persist for up to 10 years of follow-up.17 Consequently, the group of patients added to the tumor registry over the last 2 years of the study (n = 1160 [40% of the entire study population]) may have significantly reduced the observed incidence of leukemia. The three documented cases of acute myelogenous leukemia, however, did possess the cytogenetic abnormalities that are characteristic of secondary leukemia.

Kaplan et al. are to be credited for maintaining their extensive community-based hospital database. The updating of data over an adequate follow-up period, such that the duration of observation is sufficient, may prove critical to their conclusions. In addition, stratification of their analysis on the basis of accepted risk factors for secondary leukemia (e.g., cumulative cyclophosphamide and radiation therapy doses) is warranted.

In conclusion, the data presented by Kaplan et al. do not appear to be sufficiently mature to suggest the absence of an association between anthracycline use and secondary leukemia, nor do they provide convincing evidence that cyclophosphamide is not leukemogenic in patients with breast carcinoma (and particularly in those who also receive radiation therapy). Nonetheless, it is agreed that secondary leukemia is rare enough not to discourage the use of otherwise beneficial treatment options, particularly if the cumulative cyclophosamide doses administered are less than 6 g/m2. (Standard adjuvant therapy involves the administration of 2.4–3.0 g/m2, thus providing a very safe margin.) If the estimated absolute benefit of adjuvant chemotherapy is very small (i.e., < 1%),18 then the hazards and benefits of treatment should be discussed thoroughly with the patient, who should be made aware of the risk of developing secondary leukemia (particularly if radiation therapy is indicated).

In general, patients with breast carcinoma should be informed of the risk of secondary leukemia, and particularly emphasized when the benefits of adjuvant chemotherapy (which typically involves cyclophosphamide or anthracyclines) with or without radiation therapy, may not outweigh the associated risks.


  1. Top of page